Assessment of Microrna Expression in Mantle Cell Lymphoma Using High Throughput Techniques

Mantle cell lymphoma (MCL) accounts for 3–10% of all lymphomas and demonstrates a poor clinical response to current therapeutic approaches, with a median survival of 3–5 years. The natural history of MCL is heterogeneous and not well-defined by standard clinical markers as patients may die within months of diagnosis or experience long-term survival. There remains a need for reliable biomarkers of MCL prognosis. To date, global gene expression signatures have not been determined for formalin-fixed, paraffin-embedded (FFPE) MCL samples due to difficulties isolating full-length mRNA transcripts from FFPE tissues. Examining microRNA expression in FFPE samples may circumvent this problem, as this population of RNAs remains intact during processing of FFPE samples. MicroRNAs (miRs) are small, non-coding RNAs, which regulate gene expression by inhibiting mRNA translation. Although miR expression signatures have been derived for other hematological malignancies, assessment of miR expression in patients with MCL has yet to be undertaken. We hypothesize that different pathological subtypes of MCL have unique miR expression signatures, distinct from miR expression profiles of other B-cell non-Hodgkin lymphomas. Our objectives were two-fold:

1. To determine and validate miR expression in different pathological subtypes of MCL; and,

2. To compare miR expression in MCL to other B-cell non-Hodgkin lymphomas.

Total RNA was extracted from FFPE samples [17 conventional MCL, 11 blastoid MCL, 4 follicular lymphomas (Grades 1, 2, 3a, and 3b), 1 nodal marginal zone lymphoma, 1 small lymphocytic lymphoma (SLL/CLL) and 3 benign, reactive lymph nodes (normal controls)] using the RecoverAll kit for FFPE tissues (Ambion). RNA was subjected to quantitative real-time PCR (qRT-PCR) for 365 miRs and 3 endogenous control small nucleolar RNAs to obtain miR expression profiles using the TaqMan Low Density Array (TLDA) v1.0 platform (MicroFluidic card, Applied Biosystems). TLDAs were run on the ABI7900 HT analyzer with TLDA upgrade and analysed with RQ Manager software provided by Applied Biosystems. Expression profiles were correlated to pathological subtype, and hierarchical clustering, principal component analysis (PCA), and ANOVA were performed using Partek software [Partek Genomics Suite for Gene Expression Data]. Results indicate that miR expression profiles differ between B-cell, non-Hodgkin lymphoma and MCL samples. Clustering analysis and PCA both demonstrated different profiles between MCL and B-cell, non-Hodgkin lymphomas. Although PCA did not demonstrate significant differences between the conventional and blastoid MCL samples, a set of fifteen miRs may be able to distinguish these two groups of MCL, since these 15 miRs are relatively upregulated in blastoid MCL in comparison to conventional MCL. In addition, PCA revealed five (3 conventional MCL samples and 2 blastoid samples), which did not cluster with their respective groups. These five samples were from patients known to have progressive disease, indicating that such patients may have different miR expression profiles compared to patients with non-progressive disease. We conclude that high-throughput miR expression profiles can be generated from FFPE samples in B-cell non-Hodgkin lymphomas and that miR expression profiles for MCL samples differ from those of other B-cell non-Hodgkin lymphomas. Blastoid and conventional MCL samples may not have significantly differing profiles, however a set of 15 miRs appears to be able to distinguish between these two groups. Of note, samples from patients with known progressive disease have significantly different profiles from those with non-progressive disease.